District 12 Senate recount suspended

CONCORD – Secretary of State Bill Gardner suspended a state Senate recount affecting Nashua and three neighboring towns Friday night after ballot counting in Hollis came up 234 votes shy of what local officials reported on Election Day.

Another discrepancy emerged when the total number of ballots that were recounted Friday – 4,758 – was 140 less than the number local officials said actually came to the polls on Nov. 2.

“Clearly, there is an error here, but it could be easily explained once we fully check with the town clerk and moderator about what happened,” Gardner said.

Sen.-elect David Gottesman, a Nashua Democrat, defeated Hollis Republican Rep. Harry Haytayan by 188 votes, 13,316 to 13,128, in District 12.

District 12 includes Nashua Wards 1, 2, 5 and 9 along with the towns of Brookline, Hollis and Mason.

Six teams of recount volunteers poured over the ballots from two Nashua wards and Mason for eight hours Friday, which resulted in Haytayan narrowing Gottesman’s margin of victory by only 11 votes.

But the vote count in Hollis was far different than either side could have expected.

After the recount, Haytayan had 2,438 votes, or 113 less than the town reported he had gotten on Nov. 2. Gottesman had 1,942 votes, or 121 less the town concluded. A check with Hollis town officials confirmed that the seven boxes of ballots sealed after the polls closed had been sent to state archives and were delivered for the recount.

“This is just bizarre. That’s about all you can say about it at this point,” said Gottesman, who sat through much of the ballot recounting.

Haytayan spent the morning watching but later turned supervision of it to Erik Taylor, a principal in Elevare Communications of Concord that managed his campaign.

“We’re just trying to get to the bottom of these missing votes. Are there ballots that were counted on election night and aren’t before us now? We still don’t know that answer,” Taylor said.

Gottesman’s lead lawyer in the recount, Harry Judd, said whether these ballots are really missing or this discrepancy is the result of a counting error, it’s not going to change the outcome.

“The undeniable fact is the recount has showed Haytayan didn’t win his hometown by very much and that’s what the town reported on Nov. 2,’’ Judd said.

“Whether these votes are real or not, they aren’t going to change the result that Dave Gottesman won the election.”

Charles Arlinghaus, an adviser to Haytayan, said he couldn’t concede that’s the case.

“This is not something that happens very often in this state, so until we get to the bottom of it, who knows what we are dealing with?” said Arlinghaus, a former GOP state executive director.

Gardner said he would try to come up with an explanation for the inconsistency after discussions with Hollis officials.

If the recount were to resume, it won’t begin until next Thursday, he said.

Both sides reviewed a box of about 150 Hollis ballots that said “unused’’ on the front and concluded they did not count.

But Hollis Town Moderator Jim Squires told Gardner there weren’t any unused ballots that the state had given the town.

“The first thing we need to do is go back to that box and see whether all those ballots really are ‘no’ votes or if there are some valid ballots mixed in,” Gardner said.

Hollis was one of many communities on Election Day that ran out of ballots that go through the automated voting machine and had to make copies that were then hand counted.

This increases the possibility of human error, Gardner said.

If the small stack of paper ballots were counted more than once, that could have inflated the reported number of ballots cast.

The town reported 4,898 cast ballots from among 5,230 names on the checklist. This 93.6 percent turnout was nearly 3 percent higher than any other polling place in Hillsborough County on Nov. 2.

The closest towns to Hollis were Temple, which had a 90.7 percent turnout, and Mason at 90.3 percent.

If the total number of the checklist got inflated by 140, the actual turnout in Hollis would have been 90.9 percent.

What adds to the confusion is town officials reported that the number of ballots given to those who voted for president and governor were each at least 99 votes more than those voting for the state Senate.

Gardner said one possible explanation could be local officials counted the small stack of paper ballots more than once for races at the top of the ballot as they faced deadline pressure to complete those counts as soon as possible.

“All we can do is speculate now,” Gardner said. “We really don’t know.”

Assistant Attorney General Bud Fitch said Squires has the reputation as one of the state’s most thorough moderators and that he followed the vote counting instructions state officials had given him to the letter.

“He did everything in setting up this vote count that we could have wanted him to do,” he said.